We gratefully acknowledge support of this research by the NSF grant CCF-0726698.

---

People

	Evgeny Katz
	Vladimir Privman

---

Highlights

We explore experimentally, as well as model and optimize theoretically new systems that show promise for digital biochemical computing, including the first attempt for an experimental realization of error correction. Biomolecular digital Boolean logic gates, copying (fanout), error correction by utilizing redundancy, as well as signal rectification, are researched in this project. The project is based on the state-of-the-art bioelectronics and bionanotechnology advances. The experimental approach includes information processing using enzyme-based systems, encoded DNA sequences, DNAzyme biocatalyzed reactions and the use of DNA-functionalized magnetic nanoparticles. Electronic, electrochemical, and optoelectronic probes of the encoded DNA sequences and biomolecular logic gates are used to read out the results.

We have developed different biomolecular systems demonstrating Boolean logic operations OR, AND, XOR, etc., using enzymes as input signals (Fig. 2). Due to the high turnover numbers of the biocatalytic inputs, the meaningful changes in the systems were obtained in the presence of small catalytic quantities of the enzyme-inputs.

The enzyme-based logic gates were scaled-up to a logic network composed of several concatenated gates performing implication logic operations. Input signals of different physical nature were used to activate enzyme-based logic gates. Biocatalytic units associated with magnetic nanoparticles were controlled by an external variable magnetic field to generate various logic operations. This approach promises to scale-down the responsive element to a single nanoparticle functionalized with a single enzyme molecule being able to process information at the nanoscale.

To the top

---

Findings

   1. Optimization of enzyme logic gates and their circuitries
    2. Integration of enzyme logic gates with signal-responsive materials and bioelectronic devices
    3. Biomedical applications of the enzyme logic gates and their circuitries

1. Optimization of enzyme logic gates and their circuitries.

We demonstrated both experimentally and theoretically that the analog noise generation by a single enzymatic logic gate can be dramatically reduced to yield gate operation with virtually no input noise amplification. This is achieved by exploiting the enzyme's specificity when using a co-substrate that has a much lower affinity than the primary substrate. Under these conditions, we obtain a negligible increase in the noise output from the logic gate as compared to the input noise level. Experimental realizations of the AND logic gate (Figure 1, left) with the enzyme horseradish peroxidase using hydrogen peroxide and ferrocyanide as input signals resulted in low noise information processing by the enzyme gate (Figure 1, right), confirmed our general theoretical conclusions.

Figure 1. (left panel) Single-enzyme AND logic gate. (right panel) (A) Measured and (B) numerically fitted response surface for the enzymatic logic gate with ferrocyanide as one of the inputs. (C) Surface plot of the gate function quality measure, as a function of the enzyme concentration and reaction time. (D) Dependence of the gate function quality measure onon HRP concentration for different reaction times. Curves labeled (a-j) in the order indicted by the arrow correspond to different reaction times 20, 40,...,200 sec. The red dot marks our experimental conditions of [HRP](t=0)=50 nM and reaction time 60 sec.

We developed an approach aimed at optimizing the parameters of a network of biochemical logic gates (Figure 2, left) for reduction of the "analog" noise buildup. Experiments for three coupled enzymatic AND gates were reported, illustrating our procedure. Specifically, starch - one of the controlled network inputs - was converted to maltose by beta-amylase. With the use of phosphate (another controlled input), maltose phosphorylase then produced glucose. Finally, nicotinamide adenine dinucleotide (NAD+) - the third controlled input - was reduced under the action of glucose dehydrogenase to yield the optically detected signal. Network functioning was analyzed by varying selective inputs and fitting standardized few-parameters "response-surface" functions assumed for each gate (Figure 2, right). This allowed a certain probe of the individual gate quality, but primarily yielded information on the relative contribution of the gates to noise amplification. The derived information was then used to modify our experimental system to put it in a regime of a less noisy operation.

Figure 2. (left panel) Schematic representation of the enzymatic network. Upper panel: the reagents involved as inputs/outputs, and the three enzymes used as the "gate machinery," each carrying out an AND Boolean gate. The inset spells out some of the abbreviations (with additional specifications provided in the text). Lower panel: the network of three AND gates with logic-variable inputs and outputs identified. The inset gives the truth table for an AND gate. (right panel) The fitting function shown for randomly selected representative values. The two views (of the same function) illustrate the balance of the slopes of the cross-sections at output values 0 vs. 1, when one of the inputs is near 1, as well as the vanishing gradient when both inputs are 0.

To the top

2. Integration of enzyme logic gates with signal-responsive materials and bioelectronic devices.

A nanostructured system composed of enzyme-functionalized silica microparticles, ca. 74 um, and gold-coated magnetic nanoparticles, 18 +- 3 nm, modified with pH-sensitive organic shells was used to process biochemical signals and transduce the output signal into the changes of the optoelectronic properties of the assembly, Figure 3. The enzymes (glucose oxidase, invertase, esterase) covalently bound to the silica microparticles performed Boolean logic operations AND / OR processing biochemical information received in the form of chemical input signals resulting in changes of the solution pH value. Dissociation state of the organic shells on the gold-coated magnetic nanoparticles was controlled by pH changes generated in situ by the enzyme logic systems. The charge variation on the organic shells upon the reversible protonation/dissociation process resulted in the changes of the gold-layer localized surface plasmon resonance energy (LSPR), thus producing optical changes in the system. The proton transfer process allowed the functional coupling of the information processing enzyme systems with the signal transducing gold-coated magnetic nanoparticles providing their cooperative performance. Magnetic properties of the gold-coated magnetic nanoparticles allowed separation of the signal-transducing nanoparticles from the enzyme-modified signal processing silica microparticles. The reversible system operation was achieved by the Reset function returning the pH value and optical properties of the system to the initial state. This process was biocatalyzed by another immobilized enzyme (urease) activated with a biochemical signal. The studied approach opens the way to novel optical biosensors logically processing multiple biochemical signals and "smart" multi-signal responsive materials with logically switchable optical properties.

Figure 3. Optoelectronic transduction of output signals generated by the enzyme logic systems using signal-responsive nanoparticles.

Electronic transduction of biochemical signals processed by the enzyme-based OR-Reset / AND-Reset logic systems was achieved using field-effect Si-chips, Figure 4. The developed enzyme logic systems produced pH changes as the result of the biochemical reactions activated by different combinations of the chemical input signals. The signal transduction was performed by pH-responding gold nanoparticles associated with the chip interface. The transformation of the nanoparticles shells between the dissociated (negatively charged) and protonated (neutral) states was read out using capacitance-voltage or impedance spectra measurements resulting in the electronic signal that reflects the state of the system corresponding to the logic output of produced by the enzymes. The developed systems are the first examples of enzyme-based biocomputing systems interfaced with the ordinary Si-based electronics.

Figure 4. Interfacing of biomolecular computing systems with regular electronics was used to transduce logically processed biochemical signals to electronic outputs.

The modified electrode for electrocatalytic oxidation of NADH was developed using a pH-switchable redox interface, Figure 5. The operation of the modified electrode was controlled by logic operations performed by enzyme-systems processing biochemical input signals. The electrocatalytic oxidation of NADH was activated upon appropriate combinations of the signals processed by the AND / OR logic operations performed by the enzymes.

Figure 5. Logic operations AND / OR performed by the enzyme-based systems resulting in the ON and OFF states of the bioelectrocatalytic interface followed by the Reset function to complete the reversible cycle. Schematically shown cyclic voltammograms and impedance spectra correspond to the ON and OFF states of the bioelectrocatalytic electrode.

The modified interface was reset in a mute non-active state by another enzyme reaction. The coupling between the enzyme logic systems and the bioelectrocatalytic interface was achieved by pH changes produced in situ by the enzymes reactions resulting in different protonation states of the polymeric matrix associated with the electrode surface. The bioelectrocatalytic system integrated with biochemical computing systems opens the way to novel "smart" interfaces for multi-signal biosensors and signal-controlled biofuel cells. In a long perspective this approach will allow physiological control of implantable bioelectronic devices.

Further increase of the system complexity was the next aim of our research. The logic network composed of three enzymes (alcohol dehydrogenase, glucose dehydrogenase and glucose oxidase), Figure 6(A), operating in concert as four concatenated logic gates (AND/OR), was designed to process four different chemical input signals (NADH, acetaldehyde, glucose and oxygen), Figure 6(B). The cascade of biochemical reactions culminated in pH changes controlled by the pattern of the applied biochemical input signals. The "successful" set of the inputs produced gluconic acid as the final product and yielded an acidic medium, lowering the pH of a solution from its initial value of pH 6-7 to the final value of ca. 4. The pH changes produced in situ were coupled with a pH-sensitive polymer-brush-functionalized electrode, resulting in the interface switching from the OFF state, when the electrochemical reactions are inhibited, to the ON state, when the interface is electrochemically active. Soluble [Fe(CN)₆]^3-/4- was used as an external redox probe to analyze the state of the interface and to follow the changes produced in situ by the enzyme logic network, depending on the pattern of the applied biochemical signals. The chemical signals processed by the enzyme logic system and transduced by the sensing interface were read out by electrochemical means: cyclic voltammetry, Figure 7(A), and Faradaic impedance spectroscopy. The whole set of the input signal combinations included 16 variants. Those that corresponded to the logic output 1, according to the Boolean logic encoded in the logic circuitry, resulted in the acidic medium, thus activating the electrochemical process, Figure 7(B). This readout step features a "sigmoid" processing of the signals that provides "filtering" and significantly suppresses errors. Coupling between signal processing enzyme logic networks and electronic transducers will allow future "smart" bioelectronic devices to respond to immediate physiological changes and provide autonomous signaling/actuation depending on the concentration patterns of the physiological markers.

Figure 6. (A) The biocatalytic cascade used for the logic processing of the chemical input signals and producing in situ pH changes as the output signal. (B) The equivalent logic circuitry for the biocatalytic cascade.

Figure 7. (A) Sample cyclic voltammograms obtained for the ITO electrode modified with the P4VP polymer brush in: a) the initial OFF state, pH ca. 6.7, b) ON state enabled by the input combination 1,1,1,0, recorded at pH ca. 4.3, and c) in situ reset to the OFF state, pH ca. 8.8. Inset: step 1 - maximum current of the anodic feature recorded for the initial OFF state, step 2 - anodic peak current for the ON state, and step 3 - maximum current of the anodic feature after a reset to the OFF state. Deoxygenated unbuffered solution of 0.1 M Na₂SO₄, 10 mM K₃Fe(CN)₆, and 10 mM K₄Fe(CN)₆ also contained ADH, GDH, and GOx, 10 units mL^-1 each. Input A was 0.5 mM NADH, input B was 5 mM acetaldehyde, and input C was 12.5 mM glucose. Potential scan rate, 100 mV s^-1. (B) Anodic peak currents, I_p, for the 16 possible input combinations. The dotted lines show threshold values separating logic 1, undefined and logic 0 output signals.

Based on the logically controlled biocatalytic electrodes we developed more complex bioelectronic devices - biofuel cells. An enzyme-based biofuel cell with a pH-switchable oxygen electrode, controlled by enzyme logic operations processing in situ biochemical input signals, has been developed, Figure 8. Two Boolean logic gates (AND / OR) were assembled from enzyme systems to process biochemical signals and to convert them logically into pH-changes of the solution. The cathode used in the biofuel cell was modified with a polymer-brush functionalized with Os-complex redox species operating as relay units to mediate electron transport between the conductive support and soluble laccase biocatalyzing oxygen reduction. The electrochemical activity of the modified electrode was switchable by alteration of the solution pH value. The electrode was electrochemically mute at pH > 5.5 and it was activated for the bioelectrocatalytic oxygen reduction at pH < 4.5. The sharp transition between the inactive and active states was used to control the electrode activity by external enzymatic systems operating as logic switches in the system. The enzyme logic systems were decreasing the pH value upon appropriate combinations of the biochemical signals corresponding to the AND / OR Boolean logic. Then the pH-switchable electrode was activated for the oxygen reduction and the entire biofuel cell was switched ON, Figure 9. The biofuel cell was also switched OFF by another biochemical signal which resets the pH value to the original neutral value. The present biofuel cell is the first prototype of a future implantable biofuel cell controlled by complex biochemical reactions to deliver power on-demand responding in a logic way to the physiological needs.

        

            Figure 8. The biofuel cell composed of the pH-switchable logically controlled biocatalytic cathode and glucose oxidizing anode.

         

            Figure 9. The bar diagram showing the power density obtained from the biofuel cell as the logic function of different combinations of the biochemical input signals: a) OR logic gate, b) AND logic gate.

Further increase of complexity was achieved in the biofuel cell controlled by a logic network. A "smart" biofuel cell switchable ON and OFF upon application of several chemical signals processed by an enzyme logic network was designed, Figure 10(A). The biocomputing system performing logic operations on the input signals was composed of four enzymes: alcohol dehydrogenase (ADH), amyloglucosidase (AGS), invertase (INV) and glucose dehydrogenase (GDH). These enzymes were activated by different combinations of chemical input signals: NADH, acetaldehyde, maltose and sucrose. The sequence of biochemical reactions catalyzed by the enzymes models a logic network composed of concatenated AND / OR gates, Figure 10(B). Upon application of specific "successful" patterns of the chemical input signals, the cascade of biochemical reactions resulted in the formation of gluconic acid, thus producing acidic pH in the solution. This resulted in the activation of a pH-sensitive redox-polymer-modified cathode in the biofuel cell, thus, switching ON the entire cell and dramatically increasing its power output. Application of another chemical signal (urea in the presence of urease) resulted in the return to the initial neutral pH value, when the O₂-reducing cathode and the entire cell are in the mute state. The reversible activation-inactivation of the biofuel cell was controlled by the enzymatic reactions logically processing a number of chemical input signals applied in different combinations, Figure 11.

      

Figure 10. (A) The cascade of reactions biocatalyzed by alcohol dehydrogenase (ADH), amyloglucosidase (AGS), invertase (INV) and glucose dehydrogenase (GDH) and triggered by chemical input signals: NADH, acetaldehyde, maltose and sucrose added in different combinations. (B) The logic network composed of three concatenated gates and equivalent to the cascade of enzymatic reactions outlined in A.

Figure 11. V-i polarization curves obtained for the biofuel cell with different load resistances (A) and the power density as a function of the resistance load (B) and function of the current density: a) in the inactive state prior to the addition of the biochemical input signals (pH value in the cathodic compartment ca. 6), b) in the active state after the cathode was activated by changing pH to ca. 4.3 by the biochemical signals (combinations 1,1,1,0; 1,1,0,1 and 1,1,1,1), c) after the Reset function activated by the addition of 5 mM urea. Insets: Switchable i_sc (A) and power density (B) upon transition of the biofuel cell from the mute state to the active state and back performed upon biochemical signals processed by the enzyme logic network. The bar diagram (D) showing the power density produced by the biofuel cell in response to different patterns of the chemical input signals. Dash lines show thresholds separating digital 0, undefined and 1 output signals produced by the system.

The studied biofuel cells exemplify a new kind of bioelectronic devices where the bioelectronic function is controlled by a biocomputing system. Such devices will provide a new dimension in bioelectronics and biocomputing benefiting from the integration of both concepts.

Complex biochemical systems operating under control of enzyme logic with the use of switchable "smart" materials were developed at the next step of our project. Systems performing oxidative damage to biomolecules through catalytic cascades in the presence of iron-redox labile species were activated by enzyme logic gates processing chemical input signals in a logic way according to the built-in logic operations, Figure 12. AND / OR enzyme logic gates were composed of glucose oxidase (GOx) and GOx and esterase, respectively. The AND / OR logic functions of the enzyme gates were activated by application of glucose - oxygen and glucose - ethylacetate input signals, respectively. The enzyme logic gates, upon activation by specific patterns of the chemical input signals, produced acidic solutions and triggered release of redox labile iron species from a complex which is unstable under acidic conditions. This resulted in the activation of a catalytic cascade producing reactive oxygen species which yielded oxidative damage in biomolecules. Functional integration of the enzyme-based logic systems with the catalytic redox cascade performing damage in biomolecules on demand is a first step towards "smart" systems capable of programmed detection, identification, and neutralization of potential biohazards.

Figure 12. Enzyme logic gates processing chemical input signals were used to trigger release of redox active iron ions which produce reactive oxygen species in a catalytic cascade, thus resulting in oxidative damage in biomolecules.

To the top

3. Biomedical applications of the enzyme logic gates and their circuitries.

A biocomputing system composed of a combination of AND/IDENTITY logic gates based on the concerted operation of three enzymes: lactate oxidase, horseradish peroxidase and glucose dehydrogenase was designed to process biochemical information related to pathophysiological conditions originating from various injuries (Figure 13, left). Three biochemical markers: lactate, norepinephrine and glucose were applied as input signals to activate the enzyme logic system. Physiologically normal concentrations of the markers were selected as logic 0 values of the input signals, while their abnormally increased concentrations, indicative of various injury conditions were defined as logic 1 input. Biochemical processing of different patterns of the biomarkers resulted in the formation of norepi-quinone and NADH defined as the output signals. Optical and electrochemical means were used to follow the formation of the output signals for eight different combinations of three input signals (Figure 13, right). The enzymatically processed biochemical information presented in the form of a logic truth table allowed distinguishing the difference between normal physiological conditions, pathophysiological conditions corresponding to traumatic brain injury and hemorrhagic shock, and abnormal situations (not corresponding to injury). The developed system represents a biocomputing logic system applied for the analysis of biomedical conditions related to various injuries. We anticipate that such biochemical logic gates will facilitate decision-making in connection to an integrated therapeutic feedback-loop system and hence will revolutionize the monitoring and treatment of injured civilians and soldiers.

Figure 13. (left panel) Biochemical reactions catalyzed by LOx and HRP (A) and GDH (B) being used to perform AND / IDENTITY logic operation and the equivalent logic system used for processing the lactate, norepinephrine and glucose input signals (C). (right panel) Time-dependent signals corresponding to the formation of NADH generated by the combined AND-IDENTITY logic system upon application of various combinations of the input signals (glucose, lactate and NE) measured by optical means (top) and amperometrically (middle). (Bottom) Bar diagram featuring the combined AND-IDENTITY logic operation of the optical and electrochemical systems. Absorbance measurement were performed at wavelength 340 nm. Electrochemical measurements were performed at +0.75 V. Dash line shows the threshold values separating digital 0 and 1 output signals produced by the both systems.

Conclusions

The achieved results demonstrated the possibility to scale up the system complexity, to integrate biocomputing systems with optical and electronic transducers with the use of signal-responsive materials and to optimize the signal processing performed by biochemical means. The possibility of practical biomedical applications of the enzyme biocomputing systems has been demonstrated.

To the top

---

Publications

S. Vasilyev, M. Pita, E, Katz, Logic gates based on magnetic nanoparticles functionalized with a bioelectrocatalytic system, Electroanalysis, vol. 20, (2008), p. 22, DOI:10.1002/elan.200704043 Published

L. Fedichkin, E. Katz, V. Privman, Error correction and digitalization concepts in biochemical computing, J. Comput. Theor. Nanoscience, vol. 5, (2008), p. 36, DOI: 10.1166/jctn.2008.004 Published

M. Pita, E. Katz, Multiple logic gates based on electrically wired surface-reconstituted enzymes, J. Am. Chem. Soc., vol. 130, (2008), p. 36, DOI:10.1021/ja077908a Published

G. Strack, M. Ornatska, M. Pita, E. Katz, Biocomputing security system: Concatenated enzyme-based logic gates operating as a biomolecular keypad lock, J. Am. Chem. Soc., vol. 130, (2008), p. 4234, DOI:10.1021/ja7114713 Published

G. Strack, M. Pita, M. Ornatska, E. Katz, Boolean logic gates using enzymes as input signals, ChemBioChem, vol. 9, (2008), p. 1260, DOI:10.1002/cbic.200700762 Published

V. Privman, G. Strack, D. Solenov, M. Pita, E. Katz, Optimization of Enzymatic Biochemical Logic for Noise Reduction and Scalability: How Many Biocomputing Gates Can be Interconnected in a Circuit?, Proceedings of The Second International Conference on Advances in Circuits, Electronics and Micro-electronics (CENICS 2009), vol. , (2008), p. , Accepted

D. Melnikov, G. Strack, M. Pita, V. Privman, E. Katz, Analog noise reduction in enzymatic logic gates, J. Phys. Chem. B, vol. 113, (2009), p. 10472, DOI:10.1021/jp904585x Published

X. Wang, J. Zhou, T.K. Tam, E. Katz, M. Pita, Switchable electrode controlled by Boolean logic gates using enzymes as input signals, Bioelectrochemistry, vol. , (2009), p. ,     Submitted

T.K. Tam, M. Pita, E. Katz, Enzyme logic network analyzing combinations of biochemical inputs and producing fluorescent output signals: Towards multi-signal digital biosensors, Sens. Actuat. B, vol. , (2009), p. ,  DOI:10.1016/j.snb.2009.04.010 Accepted

M. Pita, T.K. Tam, S. Minko, E. Katz, Dual magneto-biochemical logic control of electrochemical processes based on local interfacial pH changes, ACS Appl. Mater. Interfaces, vol. , (2009), p. , DOI:10.1021/am900185c Accepted

V. Privman, M.A. Arugula, J. Halamek, M. Pita, E. Katz, Network analysis of biochemical logic for noise reduction and stability: A system of three coupled enzymatic AND gates, J. Phys. Chem. B, vol. 113, (2009), p. 5301, DOI:10.1021/jp810743w Published

I. Tokarev, V. Gopishetty, J. Zhou, M. Pita, M. Motornov, E. Katz, S. Minko, Stimuli-responsive hydrogel membranes coupled with biocatalytic processes, ACS Appl. Mater. Interfaces, vol. 1, (2009), p. 532, DOI:10.1021/am800251a Published

M. Motornov, J. Zhou, M. Pita, I. Tokarev, V. Gopishetty, E. Katz, S. Minko, Integrated multifunctional nanosystem from command nanoparticles and enzymes, Small, vol. 5, (2009), p. 817, DOI:10.1021/ja8076704 Published

M. Kramer, M. Pita, J. Zhou, M. Ornatska, A. Poghossian, M.J. Schoning, E. Katz, Coupling of biocomputing systems with electronic chips: Electronic interface for transduction of biochemical information, J. Phys. Chem. B, vol. 113, (2009), p. 2573, DOI:10.1021/jp808320s Published

M. Pita, S. Minko, E. Katz, Enzyme-based logic systems and their applications for novel multi-signalresponsive materials, Journal of Materials Science: Materials in Medicine, vol. 20, (2009), p. 457,  DOI:10.1007/s10856-008-3579-y Published

J. Zhou, T.K. Tam, M. Pita, M. Ornatska, S. Minko, E. Katz, Bioelectrocatylic system coupled with enzyme-based biocomputing ensembles performing Boolean logic operations: Approaching  "smart" physiologically controlled biointerfaces, ACS Appl. Mater. Interfaces, vol. 1, (2009), p. 144, DOI:10.1021/am800088d Published

M. Pita, M. Kramer, J. Zhou, A. Poghossian, M.J. Schoning, V.M. Fernandez, E. Katz, Optoelectronic properties of nanostructured ensembles controlled by biomolecular logic systems, ACS Nano, vol. 2, (2009), p. 2160, DOI: 10.1021/nn8004558 Published

M. Motornov, J. Zhou, M. Pita, V. Gopishetty, I. Tokarev, E. Katz, S. Minko,  "Chemical transformers" from nanoparticle ensembles operated with logic, Nano Lett., vol. 8, (2009), p. 2993, DOI: 10.1021/nl802059m Published

T.K. Tam, J. Zhou, M. Pita, M. Ornatska, S. Minko, E. Katz, Biochemically controlled bioelectrocatalytic interface, J. Am. Chem. Soc., vol. 130, (2009), p. 10888, DOI:10.1021/ja8043882 Published

V. Privman, G. Strack, D. Solenov, M. Pita, E. Katz, Optimization of enzymatic biochemical logic for noise reduction and scalability: How many biocomputing gates can be interconnected in a circuit?, J. Phys. Chem. B, vol. 112, (2009), p. 11777, DOI:10.1021/jp802673q Published

To the top